Clinical psychosocial programs for the rehabilitation of cocaine addicts are expensive and have high relapse rates. Numerous pharmacological agents have been used in cocaine addiction. These studies have been limited by the relatively short-term use of the pharmaceutical agent in and around the time of early cocaine abstinence, research design flaws, and in particular failure to use relapse-to-cocaine use as the major outcome variable. The dopamine agonist bromocriptine has been most widely studied and in controlled trials appears to reduce subjective distress in patients, reduce cocaine craving, and contribute to helping patients stay in their psychosocial treatment. Whether bromocriptine or any other agent reduces recidivism in cocaine dependency is not established. Recently, a new dopamine agonist, pergolide mesylate, has come on the market for the treatment of treatment-resistant Parkinson's disease. It has some similarities to bromocriptine, but in an outpatient setting for the management of cocaine dependence, this agent has some potential advantages. Pergolide is both a D1- and D2-dopamine agonist. Bromocriptine is primarily a D2-agonist and a D1-antagonist in therapeutic doses. The half-life of pergolide is sufficiently long that the drug may be administered once or at most twice daily, which will enhance compliance among addicts. Recent uncontrolled clinical data indicates that pergolide may be more effective than bromocriptine in reducing subjective craving in acutely abstinent cocaine-dependent patients. The present study seeks to recruit cocaine-dependent subjects who are enrolled in one of three outpatient addiction treatment programs, selecting males and females between ages of twenty and forty. After appropriate screening and informed consent procedures, subjects will be randomized to one of two dosages of pergolide or placebo. Subjects will be maintained on medication for twelve weeks. They will be followed off medication for an addiction sixteen weeks. The primary outcome measure will be relapse-to-cocaine use as measured by urine drug screens and collateral information from significant others. In addition to their standard psychosocial treatment, subjects will be seen weekly by a physician, nurse, and research assistant. Side effects of medication will be monitored closely.